I played tic-tac-toe against DNA

[@SteveMould]

The slowest game of noughts and crosses I've ever played. The sponsor is Incogni: The first 100 people to use code science at the link below will get 20% off: incogni.com/science Find Andrew Holding here: twitter.com/andrewholding www.twitch.tv/andrewholding instagram.com/andrewholding www.tiktok.com/@drandrewholding mastodon.social/@andrewholding kzbin.info

[@khalilahd.]

Thank you for another great video Steve

[@Reth_Hard]

1:00 Username: Challenger Password: Challenger 😛

[@SMAB2007]

This is honestly so wonderful, think of the applications! We could target certain genetic diseases such as Sickle Cell Anemia with this technology once it's developed!

[@lxlotl]

Could you post the DOI link to that paper?

[@x-crisis]

Probably the most expensive as well.

[@rentristandelacruz]

Now, DNA computing. It's only a matter of time until Steve Mould discovers slime mold computing. He can create a "Mould's Mold" computer.

[@TheCortymast]

I would love to see that!

[@FunkOsax]

😂😂

[@SomeRandomPiggo]

In British English, "Mould's Mould"

[@JessWLStuart]

I heartily approve!

[@Dziaji]

No, you di’n’t!

[@goclbert]

I've loved every episode of this series demonstrating the generalizability of computation beyond things which we usually associate with computers.

[@brodriguez11000]

Fluidics.

[@bluerizlagirl]

At the molecular level, it's really all just physics .....

[@motherisape]

Good videos for teaching a layman about computer

[@SunlightonMoon]

Well, anything can be a computer if you try hard enough

[@inferno3236]

Actually anything can be, as long as there exist two ‘states’ of a matter (called informational entropy): spin up/spin down, 0 charge/1 charge, on/off.. the theory behind all this is information theory, and this informational entropy is very similar and/or related to thermodynamic entropy.. very interesting stuff

[@CowCoder]

This guy can really make logic gates out of anything. Steve, next build a 16 bit cpu using crabs.

[@deslomator]

Crabs are a very resilient architecture, everything eventually evolves intro crabs.

[@gerardvinicombe470]

@@deslomator Ooo Ooooh ... (p.o.v. : my mind creates cognitive(?) plot points of hypothetical evolutionary convergence/divergence) " ..you take the Blue blood, the dream ends and you wake up in a rockpool breathing the dissolved oxygen.. " 🙃

[@Iyiouseismouse]

@@deslomator well… it’s more like most things tend to evolve into crab like body types 🦀

[@SedoKai]

Anyone can make logic gates out of anything. I made a basic processor out of bananas once.

[@Iyiouseismouse]

@@SedoKai yeah, but can you make a banana out of processors?

[@theadamabrams]

Here's the published paper mentioned at 3:55 www.n3cat.upc.edu/papers/Medium_Scale_Integration_of_Molecular.pdf

[@thehumblefactory]

To be fair, any time you get beat at a game, by a human, you also get beat by DNA...

[@anujmchitale]

No DNA doesn't help in real-time thinking. That's only the brain.

[@GNP3WP3W]

@@anujmchitale what is the brain made from?

[@diegotarallo2918]

@@anujmchitale which is not a static entity but a dynamic group of cells which respond to continuous adaptation. You can't rule out DNA and that's why it's a bad idea to modify your code.

[@clintonbehrends4659]

@@GNP3WP3W cells whilst the cells constuction was dictated by dna the nuerons don't calculate things using dna they use chemical recepters and electricity much like a computer Dna is not responsible for calculation

[@EEE-ro5px]

@@GNP3WP3W Neurons

[@boredwools2158]

This is the best and most intuitive lesson in microbiology I've ever seen. So good. Steve really goes all in on understanding the subject before making the video! Best educator on in the internet.

[@psi9899]

This is molecular biology

[@boredwools2158]

@@psi9899 Yes

[@thomascaldwell184]

Science communication, at its very best, explains concepts in such a way that a person not knowledgeable in that area can easily grasp it without needing to gain that detailed knowledge. The explanation uses analogies or examples clear to "laypersons." Steve is very good at that kind of science communication. I have not been clear at all on how DNA could be used to perform logic-- but Steve's explanation makes so much intuitive sense. This is what science communication is all about. Thanks Steve!

[@justforplaylists]

I think it would be good to eventually have some way to filter the output. Like, the last output triggers an enzyme to break apart a molecule so that it can fit through a pore in a membrane to go to the next stage. That way you don't need increasingly longer sequences to code complex logic.

[@jakenadalachgile1836]

great idea!

[@TehJumpingJawa]

That sounds a lot like self-modifying code. Terrifying to most programmers, and rapidly progresses beyond the comprehension of human meat bags, or even their silicone overlords. If only there was a way for DNA programs to mutate themselves, then gauge the fitness of their alteration through some kind of naturally selective mechanism...

[@justforplaylists]

@@TehJumpingJawa I get that you're joking, but also I don't think what I meant is anything like self-modifying code at all, maybe I didn't explain it well. In this video, they need a different sequence for every bit of information. So it's like if every bit in the processor needed its own unique address, and those addresses had to be stored in the BIOS. It's really inefficient. So I'm saying to split the processor into a bunch of smaller units so that each unit can reuse the same sequences. Self-modifying code is terrifying not because it will become sentient, but because it's a huge pain to debug.

[@krissp8712]

Yeah, the way you described it doesn't sound self modifying at all. It's just a way of transmitting data, the sequence isn't changing only the relevant parts would be moving.

[@misharubanov1499]

This is a great idea! And one that's been tried in a lot of different ways - for example, using hydrogel or vesicle compartments to store DNA sequences, and recycling them in other locations. Lots of great research towards this direction, happy to point you towards it!

[@bo_broadwater]

Cool! I studied strand displacement for my PhD! Never thought I'd see a Steve Mould video on it 😁

[@Sorc47]

Next goal: Make DNA run DOOM.

[@christopherlawley1842]

You mean /play/ surely?

[@whatitmeans]

@@christopherlawley1842 actually, if you can make a system that have and AND, OR, and NOT logical gates, then you can implement whatever logical instructions within it, which is called "Turing Complete", so it is able to perform whatever thing that could do a computer. So in principle, it can not only play DOOM, but even be the computer were you are playing DOOM. And amazing example of what means to be Turing Complete, its the Conway's Game of Life, a zero-sum game with just a few simple instructions, were some mathematician prove it was Turing complete, then others figure out how to implement logic gates, and now you can see on KZbin videos of the game being played within the game as it were a computer... is outstanding... and in theory same could be done with DNA: the problem now is how to implement it, because the question about If it is possible, habe been already answered with the paper results.

[@halfsourlizard9319]

@@whatitmeans It could also simulate *you*, given that there's absolutely no reason to believe that humans are computing super-Turing functions.

[@whatitmeans]

@@halfsourlizard9319 Actually is quite the opposite, you could search about how artificial networks are run on logic gates within computer servers, also, search about how actually the brain hallucinate what we call reality. Videos available for free in youtube. You could believe what you want about the meaning for them to you, but saying there is no evidence is quite against ton of facts already available.

[@halfsourlizard9319]

@@whatitmeans None of those processes are super-Turing. You're arguing the wrong point.

[@CaptainTid]

actually started laughing to myself when you explained how the AND gate physically worked, that's so insanely cool

[@matthewgiallourakis7645]

At the end of the video you mention that fluorescence isn't a useful intermediary step, but it can be! A professor that I had worked with studied modified DNA that would actuate in the presence of specific wavelengths of light. Then you could make DNA opto-isolators that could bridge DNA logic solutions without having to mix them!

[@diogenesoliveira6473]

Wow. That actually sounds like the protomolecule builders from the Expanse series: they're basically conscious light

[@plzletmebefrank]

@@diogenesoliveira6473 Not really. This is just conditionally fluorescing DNA that activates some other chemical responses in other DNA so you don't have to do quite as much pipetting. So it's basically an LED/photovoltaic/light sensor/whatever term you decide to use cause there's like 20.

[@anon_y_mousse]

That's definitely something I could get interested in. Brings me a step or two closer to building a biological computer.

[@smileyp4535]

@@plzletmebefrank but biological!

[@plzletmebefrank]

@@smileyp4535 Yep. A biological photovoltaic.

[@beeallen2743]

This is some extremely slow interfacing, but still gives you an idea of how efficient DNA is at data storage and retrieval! I mean that is the purpose of DNA but it's still pretty mind-boggling.

[@plzletmebefrank]

I'm curious now about if it'd be more stable than other data storage mediums in certain circumstances. For instance, would it be reasonable to use it as some sort of data time capsule? Though if it would be feasible to retrieve data from it without some whacko system like identifying vague chemical changes in it... You probably are living in the future with tech much better suited to that sort of purpose.

[@hyeve3551]

@@plzletmebefrank DNA data storage (and related tech) are already being researched a lot! Reading and writing custom DNA isn't especially hard with modern science, but it's very very slow.

[@anon_y_mousse]

@@plzletmebefrank Considering that it can be destroyed just by spraying alcohol on it, maybe not so robust in a static sense. Of course if you could keep it alive in some sort of medium that could be completely enclosed, maybe even encased in concrete, then we'd have something. Only problem, to have it be self maintaining, you'd need an energy source, and even radioactive sources don't last forever. I think the best ultra long term, stable data storage idea I've heard is to use giant stones to mark the data. Think Stonehenge, but with way more rocks to encode more bits.

[@coyotedomino]

@@anon_y_mousse one idea is to put important information in the genomes of entire species, in inactive genes

[@anon_y_mousse]

@@coyotedomino As long as they're hardy and not likely to go extinct that could work for a while, but mutations would occur and cause data drift. That's a weird thought, you come back to a planet a million years later and you've lost that precious photo of you and your first wife because the creatures you stored it in evolved.

[@kennymartin5976]

Thanks for the physical demos Steve! I never thought of chemical processes in mechanical terms like this before!

[@stvrdo]

Thanks for the magnificient educational videos!

[@khalilahd.]

This is actually so brilliant. It’s amazing to see how far science has come ❤

[@richbuilds_com]

This is definitely one of those videos where at the start you think: WHY? And by the end you think: Shit! The possibilities are endless! Biological computing means: Need more RAM/Processors? Ok. Give me a sec while I grow some more...

[@karolakkolo123]

@@richbuilds_com All you need to do is figure out how to make flip-flops and you can make memory. But to be fair, you would probably run out of types of molecules to be able to channel outputs to inputs

[@serotoninja7878]

The cells of the human brain are encoded by DNA. So this video shows DNA having made a game that DNA is now teaching DNA how to play. Crazy stuff.

[@WozzyA]

eh, DNA lays the groundwork for how the brain develops, but anything beyond that is the particular way your neural pathways develop due to environment

[@fadyibrahim5611]

dont forget about the DNA that the DNA is carrying in their DNA junk trunk.

[@sbiecoproductions6062]

... Which is watch all over the globe from OTHER DNA💥💥💥 also, this makes you a gaming channel @Steve?

[@trolreaper7293]

@@WozzyA that is proven incorrect your DNA tells you how to act in a situation your brain detects the situation and acts then gives feedback to the gene's depending on the situation the gene's have a chance of saying to the cell to utilize a different part of your code thus your body if the situation were to happen again uses the modified or a different gene info to react to the situation so in a way the brain makes the code while the gene's are the tools ironically

[@calebrobinson6406]

Wouldnt it be the cells that DNA says to create in the brain? Not DNA itself. Idk

[@Cyrathil]

Considering Boolean algebra kind of originated out of thought experiments of "what if we restricted ourselves to this set of operations", I am kind of interested what the algebra that comes from DNA would look like. It's interesting that X&&Y&&!Z is not just a composition of AND and NOT, so it's a separate operator. In Boolean algebra we have things like OR, XOR, NAND, etc. It'd be interesting to see DNA equivalents, or if the equivalents are impossible and what is possible.

[@3211learza]

I think just the two operators "X && Y && !Z" and "X || Y" should be enough to be touring-complete? you can emulate "X ^ Y" (where the caret is xor) by this kind of operation: (X && X && !Y) || (Y && Y && !X) and xors can combine into everything else

[@kylebowles9820]

@@3211learza you have 3 there: and &, not !, and or |. You only need two: (or | + not !) like Minecraft, or (not ! + and & a.k.a. NAND) like in real circuits.

[@kindlin]

@@kylebowles9820 The point of this thread is that the A&B&!C is (1) actual operation, as it's one physical molecule that does or does not catalyze (the way we write it right now is based on binary logic, but that's just how our language works at the moment). Now, we're only familiar with NAND and XOR gates because we've all grown up with binary computers and mostly understand binary logic, but what if DNA becomes much more common in the next few decades, maybe these kinds of more complex DNA gates will be our next set of standardized gates, with some other gate indexing scheme/acronym, like AAO gates, or DADO gates, etc.

[@Cyrathil]

@@kindlin Exactly. Under boolean algebra we have the universality of nand and xor, so what would be universal in DNA algebra. What are the extensions that might be possible under the changed ruleset?

[@3211learza]

@@Cyrathil what I was trying to say in my reply was that you can construct the equivalent of an xor gate using these DNA strands, so they should be able to combine into the same things (even if every xor gate effectively being 3 operations is kind of tedious, it should definitely work)

[@mosquitodaselva7319]

Definitely this is one of the best KZbin channel I've ever watched, this video is so awesome!

[@Yu-sj1we]

Very cool video! I am glad to see the topic of molecular computing appear in your channel. Actually, those systems you refer at the end (gates with DNA sequences as output) already exist. They can be made in many different ways (for example, look for strand displacement logic gates). It is even more interesting because there are also DNA molecules called aptazymes than can release a DNA sequence when attached to a specific molecule (any type, not just DNA). There is even a method to design these aptazymes to detect specifically any arbitrary molecule you want. This means that you can have a system that translate chemical information to DNA sequences that can be use to perform computations to give some outputs, and, eventually, perform actions based on those outputs.

[@benjaminmiller3620]

"There is even a method to design these aptazymes to detect specifically *any* arbitrary molecule"? Since when has molecular binding for more than a handful of atoms at a time been a solved problem? This is a bold claim that would have *enormous* ramifications in molecular biology, chemistry, and drug research. Can you back it up with a citation?

[@idontwantahandlethough]

@@benjaminmiller3620 i would really like that too. I want what they just said to be true but I'm a little skeptical

[@robbechristiaens6384]

Are these like riboswitches?

[@maigowang]

Man, this is totally mind-blowing!

[@AKKK1182]

Your explanation is great, the over-complication of the whole project is scary and it still seems like magic even after knowing how it works.

[@irakyl]

The first third of this video is immaculately paced, you start strongly and you answer every question as they arise in the viewer's mind

[@apollotiger]

I work on software for lab automation that works on plates just like the one you used here, and it was really neat to see this equipment being used in one of your videos! (Also, while automation isn’t necessarily applicable to a use case like yours here, it *can* help with pipetting errors - not to mention preventing RSI for people working in the labs!)

[@gammgmm]

I just love the way you present this logic gate with a model! Amazing job

[@Talkave94]

Regarding PCR - they are not really "tests" - PCR is a Polymerase Chain Reaction, they don't "test" for the presence of anything. They are just controlled reactions to create DNA products from starters which we provide it with. Making multiple products at the same time is possible with the right conditions, like multiple starters which don't react with each other, it's called multiplex PCR and it's common practice. Great video as always, very much appreciate your content Steve!

[@kevinjamesmartin7127]

If you run the result on gel electrophoresis then it is a test.

[@johnchessant3012]

Very cool! On the subject of biology and computing, there are bacterial enzymes that can figure out how to optimally undo knots in DNA via crossing switches, which is a problem mathematicians have been working on for centuries!

[@marcfruchtman9473]

The level of complexity in creating these DNA logic machines is incredible. I really admire the author(s) of this paper, "Logic gates in an Automaton". Amazing work. However, the title of the paper should probably be something more like Molecular Logic gates built from DNA" in the title or something like that. Finding this type of paper in a search by someone actually looking for this type of information will have a very difficult time finding it, unless they already knew about it.

[@Kyle-ke5fx]

Yeah literally automaton already has a specific meaning in computer science, and the title doesn't even make sense in that context.

[@amberkatanimations6585]

The "and" and "not" gate explanations were way easier to understand than i expected, yet still showing some of the complexity. Awesome job!

[@2000Meilen]

So, how long will it take for "Doom on DNA" and "Bad Apple on DNA" videos to pop up?

[@eenoogandre]

Thanks

[@majorfallacy5926]

This just shows again that in science there is an inversely proportional relationship between how useful something is vs. how cool it is

[@jaredongsing]

This video does a masterful job a decomposing an extremely complex concept into easily digestible bits

[@dmeemd7787]

This is amazing!! This channel just keeps being more and more amazing!!

[@j.jarvis7460]

I'm mind blown.... I don't even know how you have the knowledge compact this series and explain it word for word..

[@himanshukumar-xl1tj]

i am a biotechnologist and i found this information very interesting. thanks steve.

[@chetan1329]

where do you work and what do you do

[@StupidusMaximusTheFirst]

What would the applications of something like this might be? I mean, ok he emulated an operator, I suppose there's many more silly stuff we could try with DNA, but are there any applications or it's just to further knowledge? Cause really, writing programs in DNA is kinda silly I think. There wouldn't be any real world application, no-one would use DNA for this purpose in the real world. I suppose expanding our knowledge isn't as bad, even if there is no real applications. And there has been a lot of research and experimentation on your field, despite the fact we don't understand DNA, it's purpose and how it works or what it does. Scientists just prod and poke it in labs to see what would happen and that's about it. They have really generalised ideas ( and wrong) about it, they have to 'explain' something I guess, or no funding. but at this stage I really see no point other than just playing around, which is not that bad if done for the purpose of learning and safely. So the question is, do you agree or you think this would be applicable in the real world? Will DNA replace my intel x86?

[@himanshukumar-xl1tj]

@@StupidusMaximusTheFirst See, Maybe application of logic gates made from DNA will not directly find its way in computational coding, dynamic programming or even in classical programming. But it is very likely that it will be helpful in genetic engineering, allowing us to perform more controlled and sophisticated PCRs in general. I do have a degree in computer science and engineering along with biotechnology, which allows me to clearly see possible applications of this technology from very unique viewpoints. And please feel free to share contacts if you're comfortable.

[@himanshukumar-xl1tj]

@@chetan1329 I am a final year student of batchelor of technology in biotechnology and, computer science and engineering at Jaypee University of information technology H.P. solan, India. I am not working right now. Can we be friends.😊

[@StupidusMaximusTheFirst]

@@himanshukumar-xl1tj by sophisticated PCRs, I'm not sure if I understand this, I imagine like a test with a small program? Which program probably identifies specific pathogens etc? That sounds like a possible application, thanks for your answer.

[@Henipah]

Great video. That plastic DNA model is gorgeous.

[@himanshukumar-xl1tj]

I was deeply expecting transparent cross section of this mechanism as well 😅😅. Love this channel, love you Steve.😇😇😇

[@VinnieMTG2024]

Lots of Free time And ,curiously fun and entertaining! Great video!

[@dorbert3838]

Hey Steve, an interesting topic could be DNA circuitry. Might be a little difficult because it would require a lot of background information just for people to understand what the video is about, but I find it super interesting! Additionally, making 3D nanostructures with DNA is a super cool development in manufacturing technology. Would be interested to see your take on this field of research.

[@BlairdBlaird]

2:40 it's not a bad solution either, it's essentially using bitmaps for the decision tree steps, which is very easy to apply in parallel, so there are papers on bitmaps for decision trees and classifiers on GPUs.

[@theparchitect]

A few years ago I was bored during a startup at a chemical plant where I was doing the automation. During some downtime I programmed a tic tac toe game into the graphical interface. I used the same technique you used here, but the computer was a bunch faster at it than DNA, haha.

[@howard2679]

The most difficult time consuming part was pipetting all the different combinations of DNA. There is an acoustic pipetting machine that can do this in minutes rather than hours or days. This speaker is set under the tubes of DNA and fires off small drops up in the air into another tube using sound. This Labcyte Echo speaker machine can make these combinations possible without making scientists go 🤣 with pipetting work.

[@karlkastor]

I'm wondering: If you use CRISPR/CAS as a sort of search and replace function, couldn't you implement a regular expression in DNA? So you could solve any problem that a deterministic finite automaton can.

[@benjaminmiller3620]

Probably not. At least not with arbitrary DNA. First off CRISPR/CAS can't just match *any* region of DNA. It needs a specific motif to be nearby. (a PAM sequence.) Second, it doesn't actually perform an insert itself; it uses homology directed repair. That is, the cells repair mechanisms use a provided template with ends that match the break, to construct the "inserted" sequence. If you want the possibility of general replacement, you need multiple templates for what to insert. But there is no inherent mechanism to select which one will be inserted. It'll be random. Now if you use "preformatted" DNA. IE, construct a sequence with writable tokens, separated with known spacer sequences and PAMs, and only read/write those tokens with some additional mechanism to select what will be inserted/replaced , maybe you could do it. The error rate might still be a problem.

[@halfsourlizard9319]

CRISPR is strictly less expressive / powerful than regular languages / expressions.

[@mathiasplans]

All you need for it to be Turing complete is AND, NOT, and memory, so theoretically!

[@josiah42]

I use CRISPR at work and it's not a magic wand to make everything better. There are practical limitations. What they've already described here is more impressive and compact than what you could do with CRISPR. It would be interesting to see how this could be expanded to make a DNA computer involving RNA and protein complexes like topoisomerases, but I doubt Cas9 would be your first choice. RNA is nice in that it can self cleave based on sequence. This demo is a good first step towards something larger and extraordinary.

[@curtisblanco4029]

Saw your video about the ball on the turn table. I'm impressed you mathematically calculated the movement of the ball.

[@EstebanGrasso]

Those DNA probes remind me a lot of miRNAs (microRNAs, or small RNAs that are important in gene regulation). If these gates can be adapted to use those as input in a cell, I could totally see how they will be useful for research (example, if this two miRNAs are present, make the cell fluorescent, or break this other miRNA)

[@AroMaths]

Yes, exactly that! :D

[@jentezijlstra]

Exactly, interesting topic for my master thesis maybe

[@AndrewHolding]

@@jentezijlstra mRNA makes similar loops too, like in attenuation in bacteria, also riboswitches. All very cool.

[@navinsawesomeanddivineacco4221]

Dr Andrew Holding taught me a part of my module in my first year at the University of York! He's an excellent teacher and could make even the most complex of topics easy to understand! I even had the opportunity to interview him about one of his papers for a project. What a pleasant surprise seeing him in the video!

[@Eclipse-5561]

I agree. He is very good at making complex things easy to digest, even for people with no background in science, like myself

[@Hypercube9]

"We certainly didn't make them ourselves." ... My inner 12 year old self giggled at that statement. LOL

[@jentezijlstra]

This is so fricking cool. I'm studying biomedicine with a personal focus on genetics and gene technology... And I never heard of this before. So creative and cool. And they already found this like 20 years ago? I'm gonna research where we are at nowadays

[@saschaschneider9157]

"A strange game. The only winning move is not to play" --WOPR

[@timetravellingblockhead2122]

Excellent at explaining something so complicated so intuitively

[@kemoxplus]

Your videos are always awesome and show different stuff that i could never think about it 🙂

[@cameronreid8383]

You can also configure a bunch of matchboxes and some beads and get it to *learn* to play tic tac toe perfectly.

[@JessWLStuart]

Awesome concept and execution! Also, this shows why DNA computers perhaps aren't helpful compared to electronic computers - the DNA takes a long time (comparatively) to make "decisions".

[@Tehbroser]

synthetic biology and biological computation is probably at the stage that silicon computing was back in the 50s, clunky and slow, but the speed of research is insane and it’s just a matter of time before we have generalizable biological computation tools (full control of cellular functions at dna, protein, and cellular level)

[@jentezijlstra]

This could be super helpful in research, and this was already a paper from 2006! I wonder where they are now

[@douggale5962]

We could make living machines, bacteria that gobble up landfills and leave behind battery and solar panel manufacturing ingredients.

[@taktoa1]

@@Tehbroser biological computation is many many orders of magnitude worse than transistorized computers in the 50s (TRADIC in 1954 had 700 transistors ≈ 100 gates operating at 1 MHz, this is maybe 200 gates at, generously, 0.1 Hz). I think you have to go pre-Babbage to find a time when biological computation of 2022 is better than the best available computer at the time.

[@ShaLun42]

Slow, but highly parallelizable.

[@FrostCraftedMC]

now imagine what sorta crazy computation could be done when you bring transcription into the mix, and actually use the dna to create proteins that do the calculation. this is a huge step for all sorts of science and has some huge applications, tons of which we probably cant even dream of yet

[@oskaraltman]

Are you bad at noughts and crosses or something

[@Misterpirate707]

If you own the center square it's possible to guarantee that you win.

[@frag4007]

@@Misterpirate707 no its not lmao

[@mskiptr]

​@@Misterpirate707 Nah… A non-losing strategy exists for either side. If both parties play perfectly, it will always end in a tie. (see xkcd#832)

[@God-ld6ll]

how did my anticheat not detect this?

[@evanbarnes9984]

It's really easy for both sides to play a perfect game, which leads to a tie. For the games where Steve lost, he was testing the response of the system to him playing imperfectly.

[@anubhavprakash8150]

one of the best videos I have seen related to science on youtube

[@Beefster09]

Welcome to another episode of building computers with unusual things

[@kylesimpson3856]

Bro Andrew holding is my lecturer at uni this was so surreal to see him here I love guys keep it up

[@F4xP4s]

I always play a perfect game, I must have this gene!

[@mad_6519]

This is genius, and I love how it's simple enough that I can see how it works

[@Sakkura1]

You can do the same thing with RNA, and it's generally a little easier (but the RNA is less durable).

[@themultigamer5682]

Very interesting. I have an calculus teacher at my old college that recently did a study that could probably massively simplfy the logic here if the dna always starts in the middle and you take into account board symmetry.

[@Dziaji]

Congrats. You have created the best video on youtube. Grant had the lead with a few tied vids, but you have really outdone yourself on this one. Tau gang 4 life!

[@TheAdvertisement]

Can't wait for virus calculators in the future.

[@Egoistic_girl]

Are the gates still useable after the computing or are they definitely bound with the inputs? Like, does it reset?

[@markusbetts]

Ah the workhorse ABI StepOnePlus qPCR machine! Don’t know that I would be loading plates with no-replicate conditions with my bare hands though hahaha

[@ObviouslyCrap]

Are those DNA chain models generally available? What a great toy for kids and adults to learn about DNA, replication, catalysis, etc!

[@Acewey]

I would love to buy one of those DNA kits for my science class. If anyone knows where to get one, I would love to know!

[@jankonigsmark5773]

Díky!

[@ThreeEyedMonkeyMan]

"The DNA moves first, and it plays center square." I now understand how DNA beat you at Tic-Tac-Toe.

[@minerharry]

It’s interesting that the way you’re deciding which well is the DNA’s “choice” - looking at the brightest fluorescence of all the wells, using 9 simultaneous outputs - is remarkably similar to how one gets outputs from a neural net. Seems coincidental, but still cool!

[@wolvenar]

Correct me if I'm wrong but, can't this same thing be done with fairly basic chemistry. I have seen and/or gates done with various chemicals before.

[@kellymoses8566]

Dna is cooler

[@jibcot8541]

Yes I think you are correct but DNA is just chemical as well adenine (A), guanine (G), cytosine (C), and thymine (T) but I guess it's cool that it is also the instruction manual of all know life including us.

[@wolvenar]

Oh no doubt doing it with DNA is cooler.

[@goclbert]

People have gone further than that. People have created circuits using enzymes which can perform operations like exponentiation, multiplication, addition, subtraction, logarithms and roots. We also have level detectors, oscillators and switches. In a more general sense, we're just doing things biology already does.

[@crispisauce]

That's one expensive game of tic tac toe!

[@evanbarnes9984]

Just putting in a plug for The Thought Emporium! Great channel to check out if you want to learn more about the process of creating and working with little snippets of custom DNA

[@brunobucciarati4792]

I feel like the next step after this is running doom off of a string of dna

[@LARVAMOLT]

i love that the DNA "cheated" lol

[@AndrewHolding]

It did... it wasn't Steve's pipetting at all.

[@LARVAMOLT]

@@AndrewHolding we need a DNA vs DNA game!!

[@keyboardtalk]

DNA is Turing complete

[@ThisCanBePronounced]

Boolean logic, DNA, Tic tac toe?? What is Steve Mould up to this time??

[@KonstantinDankov]

This really blew my mind! Awesome stuff!

[@vennic]

The biggest wonder here is how humans continue to find games with such limitation entertaining

[@guilhermeradloff]

Next up: Can we run Doom on DNA?

[@animalpeeps]

Oh fun! I used to use qPCR in me previous job, and it was pretty gratifying getting good data/plots, and knowing my samples were done successfully. Used the same program too! Always had to be careful with working with DNA/RNA, but it was kind of fun

[@alessioulivi6734]

This was amazing, seriously

[@zombiesbyte331]

Hey Steve, I went looking for one of your vids on youTube and ended up just listening to you sing "Let Her Go", great song, you're a man of many talents ;D

[@WrappedBaggage]

I'm currently working on a side project at the moment, making a tic-tac-toe game in python for practice. Weird how my favorite content creators have a tendency to feature whatever I'm doing at the moment in their content. Great video!

[@TechnoL33T]

They might beat YOU at tic tac toe, but I'll never lose this solved game.

[@学数学-w6k]

Dear Steve, I d love to see you making a video explaining the nature of wave. Or, consider a simplest setup, such that a certain event act with wave property (e.g. diffraction). Then we can understand the necessary condition to create a wave-like phenomenon. For example, consider a chain of beans in gravity-free environment, we hit the first bean with a force (try different direction) and observe. Later consider a 2-dimensional web of beans, then hit one bean with a force normal to the plane and observe that the displacement transverse, we can place a half wall (resp. Double slit) at a distance and observe that diffraction (resp. Interference ) happen. I guess a wave-like phenomenon happen whenever, 1) in a evenly distributed space, 2) a change of certain property at a point x would cause the neighbors to have a similar change (can be the same if ideal, but would be weaker in reality due to friction).

[@bluerizlagirl]

So if the quencher robs energy from the fluorescent molecule ..... ..... Does that make it a Joule thief?

[@waclawkoscielniak9291]

Steve, maybe you and Andrew could try to run an algorithm that I came up with. My claim is that a quantum computer cannot run it satisfactorily. Perhaps a DNA computer can handle it. 1) x=0 2) Determine x 3) Save x, don't overwrite the old value of x 4) Repeat steps 1, 2, and 3 a reasonably large number of times. 5) Plot distribution of x If a computer is accurate, it would be just a vertical line. Any other device will give some strange distribution. DNA device, who knows?

[@martinellis38]

When I was little, in the 70s there was in Melbourne Museum a machine that you could play noughts and crosses against. You would also only loose and draw. This DNA processor reminds me of that. I don't think it's there any more, even in the 70s it was getting less impressive against modern technology. Also XKCD has the perfect set of moves in cartoon form

[@she_be_rosie2419]

i just saw a similar topic in a lecture a few weeks ago - that group used dna to implement a (small) neural network to run the mnist digit recognition among other things - very cool hehe

[@robinsonmedia6370]

Magnificent presentation!

[@Neoentrophy]

Selecting the centre square is a significant advantage in tic-tac-toe, you can at least claim you were playing with one hand tied behind your back

[@lukefuller284]

Next up: Steve Mould runs Doom on DNA computer simulated by water computer

[@TheArcSet]

Thanks for this. A really amazing explanation.

[@tangjtl499]

it's kinda weird but feeling some nostalgia for my masters dissertation this summer watching this haha, working with DNA, using qPCR and the stepone machine/software

[@AndrewHolding]

That machine actually is famous in it's own right, it was the first machine to be donated by a University to a Hopsital during the start of the pandemic. It was used to test staff samples.

[@ShALLaX]

This is incredibly well explained.

[@tango_doggy]

This really reminds me of how one of the first electrical ai's was made to play a perfect game of tic tac toe.

[@potaatobaked7013]

I think it'd be interesting to embed randomized and gates of DNA in a bacteria and let it mutate, then play tic tac toe and find a way of scoring how close they were to winning, take the highest scorers then put them back in the bacteria to mutate. It'd be like artificial intelligence but biology

[@regulate.artificer_g23.mdctlsk]

A *bacterium ("bacteria" is plural)

[@potaatobaked7013]

@@regulate.artificer_g23.mdctlsk bacterium is the singular for an individual organism, bacteria is the singular for a strain and the plural for organism

[@regulate.artificer_g23.mdctlsk]

@@potaatobaked7013 oh, okay. I forgot about that.